In recent years, small unmanned aircraft systems (sUAS) have been used widely to monitor animals because of their customizability, ease of operating, ability to access difficult to navigate places, and potential to minimize disturbance to animals. Automatic identification and classification of animals through images acquired using a sUAS may solve critical problems such as monitoring large areas with high vehicle traffic for animals to prevent collisions, such as animal-aircraft collisions on airports. In this research we demonstrate automated identification of four animal species using deep learning animal classification models trained on sUAS collected images. We used a sUAS mounted with visible spectrum cameras to capture 1288 images of four different animal species: cattle (Bos taurus), horses (Equus caballus), Canada Geese (Branta canadensis), and white-tailed deer (Odocoileus virginianus). We chose these animals because they were readily accessible and white-tailed deer and Canada Geese are considered aviation hazards, as well as being easily identifiable within aerial imagery. A four-class classification problem involving these species was developed from the acquired data using deep learning neural networks. We studied the performance of two deep neural network models, convolutional neural networks (CNN) and deep residual networks (ResNet). Results indicate that the ResNet model with 18 layers, ResNet 18, may be an effective algorithm at classifying between animals while using a relatively small number of training samples. The best ResNet architecture produced a 99.18% overall accuracy (OA) in animal identification and a Kappa statistic of 0.98. The highest OA and Kappa produced by CNN were 84.55% and 0.79 respectively. These findings suggest that ResNet is effective at distinguishing among the four species tested and shows promise for classifying larger datasets of more diverse animals.

The southeastern U.S. is widely known as a bastion of privately-owned, managed pine (Pinus spp.) forests, comprised primarily of native pine species. The region supports high levels of biodiversity, but also a multi-billion-dollar forest products economy critical to socioeconomic stability of rural areas. We conducted a systematic review of studies focused exclusively on avifaunal associations within privately-owned, managed pine landscapes in the southeastern U.S. We used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis framework to examine all available studies that evaluated aspects of avian diversity, abundance, and community composition across a suite of forest types, stand ages, and forest management practices within southeastern managed pine systems in the last 70 years. We screened 160 records through primary database searches, and 1696 secondary records from supplemental searches and other sources, and identified 103 relevant articles for inclusion. As expected, although there is no single forest management practice that best provides for avian communities, we found practices that: (1) involve prudent site preparation; (2) promote forest thinning and intermediate management practices; (3) provide non-pine vegetative cover; (4) supply fine- and meso-filter resources such as retained snags and coarse woody debris; and (5) promote heterogeneity in cover types, largely enhanced value of southeastern managed pine systems to avian communities. Overall, it appears that avian communities can be best maintained by providing a diverse mosaic of forest conditions in managed pine landscapes. Key research gaps include improving understanding avian population demographics, such as survival, reproduction, and dispersal.

Background: Small unmanned aircraft systems (sUAS) are replacing or supplementing manned aircraft and ground-based surveys in many animal monitoring situations due to better coverage at finer spatial and temporal resolutions, access, cost, bias, impacts, safety, efficiency, and logistical benefits. Various sUAS models and sensors are available with varying features and usefulness depending on survey goals. However, justification for selection of sUAS and sensors are not typically offered in published literature and existing reviews do not adequately cover past and current sUAS applications for animal monitoring nor their associated sUAS model and sensor technologies, taxonomic and geographic scope, flight conditions and considerations, spatial distributions of sUAS applications, and reported technical difficulties. We outline a systematic map protocol to collect and consolidate evidence pertaining to sUAS monitoring of animals. Our systematic map will provide a useful synthesis of current applications of sUAS-animal related studies and identify major knowledge clusters (well-represented subtopics that are amenable to full synthesis by a systematic review) and gaps (unreported or underrepresented topics that warrant additional primary research) that may influence future research directions and sUAS applications. Methods: Our systematic map will investigate the current state of knowledge using an accurate, comprehensive, and repeatable search. We will find relevant peer-reviewed and grey literature as well as dissertations and theses using online publication databases, Google Scholar, and by request through a professional network of collaborators and publicly available websites. We will use a tiered approach to article exclusion with eligible studies being those that monitor (i.e., identify, count, estimate, etc.) terrestrial vertebrate animals. Extracted data concerning sUAS, sensors, animals, methodology, and results will be recorded in Microsoft Access. We will query and catalogue evidence in the final database to produce tables, figures, and geographic maps to accompany a full narrative review that answers our primary and secondary questions.

Conservation prioritization frameworks are used worldwide to identify species at greatest risk of extinction and to allocate limited resources across regions, species, and populations. Conservation prioritization can be impeded by ecological knowledge gaps and data deficiency, especially in freshwater species inhabiting highly complex aquatic ecosystems. Therefore, we developed a flexible approach that calculates a species' imperilment risk based on the conservation principles of resiliency, redundancy, and representation (i.e., the “three R's”). Our approach organizes data on species traits, distributions, population connectivity, and threats within a Bayesian belief network capable of predicting resiliency and redundancy within representative ecological settings. Empirical data and expert judgment inform the model to provide robust and repeatable risk assessments for rare and data‐deficient species. The model calculates resiliency at hierarchical spatial scales from distributional trends and population strength. Redundancy is estimated from the connectivity and quantities of extant populations. Resiliency, redundancy, and species' inherent vulnerability based on species traits collectively estimate extirpation risk within each unique ecological setting. Extirpation risks across ecological settings characterize representation and are aggregated to estimate global imperilment risk. We demonstrate the model's utility with Piebald Madtom (Noturus gladiator), a species petitioned for listing under the U.S. Endangered Species Act. Our results revealed that resiliency, redundancy, and extirpation risks can vary spatially across the species' range while identifying populations where additional sampling could disproportionally reduce uncertainty in estimated global imperilment risk. Our approach could standardize and expedite conservation status assessments, identify opportunities for early management intervention of at‐risk species and populations, and strategically reduce uncertainty by focusing monitoring and research on priority information gaps.

Strategic, data driven conservation approaches are increasing in popularity as conservation communities gain access to better science, more computing power, and more data. High resolution geospatial data, indicating ecosystem functions and economic activity, can be very useful for any conservation expert or funding agency. A framework was developed for a data driven conservation prioritization tool and a data visualization tool. The developed tools were then implemented and tested for the U.S. Gulf of Mexico coastal region defined by the Gulf Coast Ecosystem Restoration Council. As a part of this tool development, priority attributes and data measures were developed for the region through 13 stakeholder charrettes with local, state, federal, and other non-profit organizations involved in land conservation. This paper presents the measures that were developed to reflect stakeholder priorities. These measures were derived from openly available geospatial and non-geospatial data sources. This database contained 19 measures, aggregated into a one km² hexagonal grid and grouped by the overarching goals of habitat, water quality and quantity, living coastal and marine resources, community resilience, and economy. The developed measures provided useful data for a conservation planning framework in the U.S. Gulf of Mexico coastal region.